Path lookups that traverse encrypted symlink(s) are very slow because
each encrypted symlink needs to be decrypted each time it's followed.
This also involves dropping out of rcu-walk mode.
Make encrypted symlinks faster by caching the decrypted symlink target
in ->i_link. The first call to fscrypt_get_symlink() sets it. Then,
the existing VFS path lookup code uses the non-NULL ->i_link to take the
fast path where ->get_link() isn't called, and lookups in rcu-walk mode
remain in rcu-walk mode.
Also set ->i_link immediately when a new encrypted symlink is created.
To safely free the symlink target after an RCU grace period has elapsed,
introduce a new function fscrypt_free_inode(), and make the relevant
filesystems call it just before actually freeing the inode.
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
->lookup() in an encrypted directory begins as follows:
1. fscrypt_prepare_lookup():
a. Try to load the directory's encryption key.
b. If the key is unavailable, mark the dentry as a ciphertext name
via d_flags.
2. fscrypt_setup_filename():
a. Try to load the directory's encryption key.
b. If the key is available, encrypt the name (treated as a plaintext
name) to get the on-disk name. Otherwise decode the name
(treated as a ciphertext name) to get the on-disk name.
But if the key is concurrently added, it may be found at (2a) but not at
(1a). In this case, the dentry will be wrongly marked as a ciphertext
name even though it was actually treated as plaintext.
This will cause the dentry to be wrongly invalidated on the next lookup,
potentially causing problems. For example, if the racy ->lookup() was
part of sys_mount(), then the new mount will be detached when anything
tries to access it. This is despite the mountpoint having a plaintext
path, which should remain valid now that the key was added.
Of course, this is only possible if there's a userspace race. Still,
the additional kernel-side race is confusing and unexpected.
Close the kernel-side race by changing fscrypt_prepare_lookup() to also
set the on-disk filename (step 2b), consistent with the d_flags update.
Fixes: 28b4c26396 ("ext4 crypto: revalidate dentry after adding or removing the key")
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Plaintext dentries are always valid, so only set fscrypt_d_ops on
ciphertext dentries.
Besides marginally improved performance, this allows overlayfs to use an
fscrypt-encrypted upperdir, provided that all the following are true:
(1) The fscrypt encryption key is placed in the keyring before
mounting overlayfs, and remains while the overlayfs is mounted.
(2) The overlayfs workdir uses the same encryption policy.
(3) No dentries for the ciphertext names of subdirectories have been
created in the upperdir or workdir yet. (Since otherwise
d_splice_alias() will reuse the old dentry with ->d_op set.)
One potential use case is using an ephemeral encryption key to encrypt
all files created or changed by a container, so that they can be
securely erased ("crypto-shredded") after the container stops.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Close some race conditions where fscrypt allowed rename() and link() on
ciphertext dentries that had been looked up just prior to the key being
concurrently added. It's better to return -ENOKEY in this case.
This avoids doing the nonsensical thing of encrypting the names a second
time when searching for the actual on-disk dir entries. It also
guarantees that DCACHE_ENCRYPTED_NAME dentries are never rename()d, so
the dcache won't have support all possible combinations of moving
DCACHE_ENCRYPTED_NAME around during __d_move().
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Make various improvements to fscrypt dentry revalidation:
- Don't try to handle the case where the per-directory key is removed,
as this can't happen without the inode (and dentries) being evicted.
- Flag ciphertext dentries rather than plaintext dentries, since it's
ciphertext dentries that need the special handling.
- Avoid doing unnecessary work for non-ciphertext dentries.
- When revalidating ciphertext dentries, try to set up the directory's
i_crypt_info to make sure the key is really still absent, rather than
invalidating all negative dentries as the previous code did. An old
comment suggested we can't do this for locking reasons, but AFAICT
this comment was outdated and it actually works fine.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Currently, trying to rename or link a regular file, directory, or
symlink into an encrypted directory fails with EPERM when the source
file is unencrypted or is encrypted with a different encryption policy,
and is on the same mountpoint. It is correct for the operation to fail,
but the choice of EPERM breaks tools like 'mv' that know to copy rather
than rename if they see EXDEV, but don't know what to do with EPERM.
Our original motivation for EPERM was to encourage users to securely
handle their data. Encrypting files by "moving" them into an encrypted
directory can be insecure because the unencrypted data may remain in
free space on disk, where it can later be recovered by an attacker.
It's much better to encrypt the data from the start, or at least try to
securely delete the source data e.g. using the 'shred' program.
However, the current behavior hasn't been effective at achieving its
goal because users tend to be confused, hack around it, and complain;
see e.g. https://github.com/google/fscrypt/issues/76. And in some cases
it's actually inconsistent or unnecessary. For example, 'mv'-ing files
between differently encrypted directories doesn't work even in cases
where it can be secure, such as when in userspace the same passphrase
protects both directories. Yet, you *can* already 'mv' unencrypted
files into an encrypted directory if the source files are on a different
mountpoint, even though doing so is often insecure.
There are probably better ways to teach users to securely handle their
files. For example, the 'fscrypt' userspace tool could provide a
command that migrates unencrypted files into an encrypted directory,
acting like 'shred' on the source files and providing appropriate
warnings depending on the type of the source filesystem and disk.
Receiving errors on unimportant files might also force some users to
disable encryption, thus making the behavior counterproductive. It's
desirable to make encryption as unobtrusive as possible.
Therefore, change the error code from EPERM to EXDEV so that tools
looking for EXDEV will fall back to a copy.
This, of course, doesn't prevent users from still doing the right things
to securely manage their files. Note that this also matches the
behavior when a file is renamed between two project quota hierarchies;
so there's precedent for using EXDEV for things other than mountpoints.
xfstests generic/398 will require an update with this change.
[Rewritten from an earlier patch series by Michael Halcrow.]
Cc: Michael Halcrow <mhalcrow@google.com>
Cc: Joe Richey <joerichey@google.com>
Signed-off-by: Eric Biggers <ebiggers@google.com>
Use a common function for fscrypt warning and error messages so that all
the messages are consistently ratelimited, include the "fscrypt:"
prefix, and include the filesystem name if applicable.
Also fix up a few of the log messages to be more descriptive.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
gcc versions prior to 4.6 require an extra level of braces when using a
designated initializer for a member in an anonymous struct or union.
This caused a compile error with the 'struct qstr' initialization in
__fscrypt_encrypt_symlink().
Fix it by using QSTR_INIT().
Reported-by: Andrew Morton <akpm@linux-foundation.org>
Fixes: 76e81d6d50 ("fscrypt: new helper functions for ->symlink()")
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Filesystems don't need fscrypt_fname_encrypted_size() anymore, so
unexport it and move it to fscrypt_private.h.
We also never calculate the encrypted size of a filename without having
the fscrypt_info present since it is needed to know the amount of
NUL-padding which is determined by the encryption policy, and also we
will always truncate the NUL-padding to the maximum filename length.
Therefore, also make fscrypt_fname_encrypted_size() assume that the
fscrypt_info is present, and make it truncate the returned length to the
specified max_len.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Currently, when encrypting a filename (either a real filename or a
symlink target) we calculate the amount of NUL-padding twice: once
before encryption and once during encryption in fname_encrypt(). It is
needed before encryption to allocate the needed buffer size as well as
calculate the size the symlink target will take up on-disk before
creating the symlink inode. Calculating the size during encryption as
well is redundant.
Remove this redundancy by always calculating the exact size beforehand,
and making fname_encrypt() just add as much NUL padding as is needed to
fill the output buffer.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Filesystems also have duplicate code to support ->get_link() on
encrypted symlinks. Factor it out into a new function
fscrypt_get_symlink(). It takes in the contents of the encrypted
symlink on-disk and provides the target (decrypted or encoded) that
should be returned from ->get_link().
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Currently, filesystems supporting fscrypt need to implement some tricky
logic when creating encrypted symlinks, including handling a peculiar
on-disk format (struct fscrypt_symlink_data) and correctly calculating
the size of the encrypted symlink. Introduce helper functions to make
things a bit easier:
- fscrypt_prepare_symlink() computes and validates the size the symlink
target will require on-disk.
- fscrypt_encrypt_symlink() creates the encrypted target if needed.
The new helpers actually fix some subtle bugs. First, when checking
whether the symlink target was too long, filesystems didn't account for
the fact that the NUL padding is meant to be truncated if it would cause
the maximum length to be exceeded, as is done for filenames in
directories. Consequently users would receive ENAMETOOLONG when
creating symlinks close to what is supposed to be the maximum length.
For example, with EXT4 with a 4K block size, the maximum symlink target
length in an encrypted directory is supposed to be 4093 bytes (in
comparison to 4095 in an unencrypted directory), but in
FS_POLICY_FLAGS_PAD_32-mode only up to 4064 bytes were accepted.
Second, symlink targets of "." and ".." were not being encrypted, even
though they should be, as these names are special in *directory entries*
but not in symlink targets. Fortunately, we can fix this simply by
starting to encrypt them, as old kernels already accept them in
encrypted form.
Third, the output string length the filesystems were providing when
doing the actual encryption was incorrect, as it was forgotten to
exclude 'sizeof(struct fscrypt_symlink_data)'. Fortunately though, this
bug didn't make a difference.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Introduce a helper function which prepares to look up the given dentry
in the given directory. If the directory is encrypted, it handles
loading the directory's encryption key, setting the dentry's ->d_op to
fscrypt_d_ops, and setting DCACHE_ENCRYPTED_WITH_KEY if the directory's
encryption key is available.
Note: once all filesystems switch over to this, we'll be able to move
fscrypt_d_ops and fscrypt_set_encrypted_dentry() to fscrypt_private.h.
Acked-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Introduce a helper function which prepares to rename a file into a
possibly encrypted directory. It handles loading the encryption keys
for the source and target directories if needed, and it handles
enforcing that if the target directory (and the source directory for a
cross-rename) is encrypted, then the file being moved into the directory
has the same encryption policy as its containing directory.
Acked-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Introduce a helper function which prepares to link an inode into a
possibly-encrypted directory. It handles setting up the target
directory's encryption key, then verifying that the link won't violate
the constraint that all files in an encrypted directory tree use the
same encryption policy.
Acked-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Add a helper function which prepares to open a regular file which may be
encrypted. It handles setting up the file's encryption key, then
checking that the file's encryption policy matches that of its parent
directory (if the parent directory is encrypted). It may be set as the
->open() method or it can be called from another ->open() method.
Acked-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Eric Biggers